As the world of AI agents advances, so does the need for these intelligent assistants to seamlessly integrate with existing enterprise systems. At MuleSoft, we’re excited to be at the forefront of this transformation, particularly with the growing momentum around the Model Context Protocol (MCP) and our capabilities that bring actionability to the agents.
We’ll explain how you can leverage MuleSoft’s capabilities to:
- Create MCP servers using natural language processing (NLP) via Anypoint Code Builder’s (ACB) Cursor extensions
- Generate or modify flows (Grounded on MuleSoft curated examples)
- Use hundreds of MuleSofts out-of-the-box connectors to common enterprise applications for your actionability use cases
- Run and debug locally
- Deploy applications to CloudHub 2.0 or Runtime Fabric
The rise of Model Context Protocol (MCP) and agentic integration
Model Context Protocol (MCP) is rapidly emerging as a critical industry standard, providing a standardized way for applications to offer context, tools, and prompts to large language models (LLMs). This open protocol is designed to standardize how AI models connect to diverse data sources and tools, fostering greater interoperability across systems.
Organizations are striving to build applications that can be interacted with in natural language while surfacing data from and taking action in a multitude of systems across a company’s key domains. Now imagine if you can build such an application using the use cases and a series of prompts rather than the traditional development tools and methods.
Build MCP servers with natural language using Anypoint Code Builder’s Cursor extension
One of the most exciting developments in this space is the ability to generate MuleSoft applications that function as MCP servers using natural language prompts within an IDE. Our focus in the initial release of this new approach will be a Cursor-based experience.
Imagine a scenario where you can simply tell your IDE: “I want to create an mulesoft MCP Server containing Anypoint MCP and Twilio Connectors, with operations like create call, update call, Update call feedback,” and have a MuleSoft application generated for you, which pending some configuration changes can be run locally, tested, and then deployed to CH2.
This is precisely what MuleSoft is enabling with a new AI-integrated server that brings MuleSoft’s development and deployment capabilities to any IDE that supports MCP.
For more information, read how to Get Started With Extending MuleSoft to AI-Native IDEs.
Extensibility use cases and agentic experiences
By easily creating MCP servers from MuleSoft connectors, you can unlock a wide range of extensibility use cases for your AI agents:
- Connecting to enterprise systems using Anypoint Connectors: Empower your agents to interact with critical business applications like Salesforce, SAP, Workday, and more, by exposing their operations as MCP tools.
- Automating business processes: Enable agents to trigger complex workflows and orchestrate tasks across disparate systems. For example, an agent could create a ticket in Zendesk, update an investigation in Jira, or retrieve Oracle ERP details like leave balances.
- Enriching agent context: Provide agents with real-time, structured data from various sources by exposing resources through MCP, allowing for more informed decision-making.
Use cases for Mule MCP servers generated from connectors
Here are a few sample use cases that can be fulfilled by Mule MCP Servers generated using NLP:
- Employee leave balance inquiry: An employee agent leveraging an MCP server, generated from a Anypoint connector to an Oracle or Workday systems, retrieves an employee’s leave balance in response to a natural language query like: “What’s my current leave balance?” or schedules vacation upon receipt of a prompt like: “Apply annual leave from July 10–14, 2025”
- Jira ticket status check: An agent in a customer service context could query a Jira MCP server (generated from the Anypoint Jira connector) to check the status of a specific ticket based on a customer’s request.
- ServiceNow incident creation: An IT support agent could, through an MCP server generated from the Anypoint ServiceNow connector, create a new incident in ServiceNow based on a user’s problem description.
These examples demonstrate the power of combining MuleSoft’s extensive connector library with the simplicity of NLP-driven MCP server generation.
Empowering AI agents with NLP-generated MCP servers is a game-changer for enterprise integration. MuleSoft is making it easier than ever to connect your intelligent assistants to critical business systems, unlocking new levels of automation and insight. Get ready to transform your agentic experiences with seamless, natural language-driven connectivity.
